6 ms·

I find it amusing that so many people is skeptic about general AI, trying to come up with arguments that it will never happen, while at the same time we do not

• bamboozled 4y ago
  Well then it's funny to think we can build something we don't even have a good definition for? Sounds almost a bit insane?
  • hexomancer 4y ago
    We didn't know what electricity really was when we created electric lightbulbs.
    • timid_oshima 4y ago
      Hadn’t it been thoroughly studied and understood for over 100 years at that point? Ben Franklin was studying it, after all.
      • hexomancer 4y ago
        Electron was discovered 1897 but the first lightbulb was 1802. There are more examples as well, as another user has commented, we didn't know what fire is until relatively recently but we have been using it for thousands of years.
        • timid_oshima 4y ago
          That sounds a bit early - whats your source? This one says at least mid 19th century for not practical electric lights https://www.energy.gov/articles/history-light-bulb https://www.energy.gov/articles/history-light-bulb But either way, is knowing the electron knowing electricity? There are so many properties of it that can be known and manipulated without that insight- and indeed they built up that understanding to reach practical engineering and use of electricity. That’s what I think is being gotten at wrt intelligence. “Knowing” something isn’t necessarily about being aware of its smaller parts.
          • hexomancer 4y ago
            > That sounds a bit early - whats your source? The source you linked already cites it as the first arc lamp. > But either way, is knowing the electron knowing electricity? There are so many properties of it that can be known and manipulated without that insight- and indeed they built up that understanding to reach practical engineering and use of electricity. That’s what I think is being gotten at wrt intelligence. Yes. That is exactly my point. We don't need to entirely understand what intelligence is in order to be able to create it. The same way we didn't know what fire is, but we created it with no problem.
            • timid_oshima 4y ago
              Ok, I’ll give you the lightbulb point :). But we can hardly define intelligence, let alone “entirely understand” it. A child could give a good , practical definition of fire and manipulate it skillfully thousands of years ago. Not so much us grown adults wrt intelligence today.
              • hexomancer 4y ago
                Here is a child-like, practical definition of intelligence: "ability to solve problems".
                • timid_oshima 4y ago
                  That’s a good start, but doesn’t help us since even the most basic programs can fit that definition and are not what we mean when we say AGI. I have yet to see anything approaching a useful definition of intelligence across several discussions of impressive new language, and other statistical models - it feels like we should have that before talking about making it real! With fire, even “when things turn from themselves to ashes and produce heat” (which I imagine a prehistoric child could come up with) distinguishes fire usefully from most other phenomena in the world
                  • mannykannot 4y ago
    • dinkumthinkum 4y ago
      That analogy doesn’t work. We knew we were trying to build a light bulb. There were properties of electricity, a complex physical phenomena, that we did not understand. However, we have a rigorous understanding of Turing machines. We have a nascent understanding of human intelligence.
      • hexomancer 4y ago
        We don't have a rigorous understanding of neural networks.
  • estevaoam 4y ago
    Sorry but your argument doesn’t make sense. Engineering/building is different than understanding how things works. Surely is related: if you build something you usually can learn how things work - and vice-versa. Insane to me is to affirm that something has a limit when you don’t fully understand it.
    • timid_oshima 4y ago
      What are other examples of things we engineer before knowing what they are? I’m honestly having trouble thinking of any. And I think it’s a shame we haven’t considered what the limits are on “intelligence” very much. Knowing them has been immeasurably valuable in software engineering and algorithms, for one.
      • estevaoam 4y ago
        I agree with you. I also think it is useful to understand limits. This is not what I’m arguing against, it is affirming limits without the knowledge to do so. Until we don’t really figure out how our intelligence works we can’t say we can’t reproduce it algorithmically. We can find plenty of examples where engineering advanced beyond our understanding: bridges, boats, tables, etc. When humanity started building those we didn’t have the full picture of our physics yet we build and used those for more than 1000s years. Again, I’m not saying we can’t engineer something from knowledge, I’m saying we can’t affirm limits onto something we don’t fully understand yet. IMHO this paper is speculative and biased towards our human intelligence. From all the advances I have seen so far in the AI landscape, I’m growing more and more skeptic that our intelligence is something so complex that we can’t replicate.
      • inkblotuniverse 4y ago
        We built airplanes before really understanding biological flight (especially e.g. bees).
  • nohat 4y ago
    Mog: “What truly is fire? The divine blessing stolen by Prometheus? Concentrated Phlogiston? The element of change? Is it not madness to seek to create something we don’t even have a good definition of?” Grog: “Grog rubs two sticks together” Lowers voice and looks around furtively “really hard.”
    • timid_oshima 4y ago
      Grog could define fire in a very real way even back then, which is how he knew he’d so easily created it . It is hard for us to even know intelligence (the one we mean in AGI) when we see it, much less create it, no?
      • nohat 4y ago
        We can define intelligence in a very real, practical way now. We see and identify intelligence all the time in humans and in animals and in AI. We may not be perfect at identifying it (just like grog might mistake a rising sun for a forest fire), but we don't need a perfect mathematical or philosophical definition that we all agree on to create it. We just need to rub sticks together really hard.
        • timid_oshima 4y ago
          The person you originally replied to and I disagree that we have any real, practical definition. I can recognize what humans and to an extent animals do as intelligent, but haven’t seen a definition that separates that intelligent behavior from them. I have never seen anything that’s been called ai do something I could call intelligent in that animal-like sense (though some have been impressive in the same way Google / page rank was impressive when it first came out) So, I don’t see why rubbing these statistical model sticks should suddenly burst into intelligence, but I’m open to seeing convincing reasoning on that at some point. I wouldn’t invest time or energy in the meantime and like that original poster, think it’s kinda insane to if my goal was to see human-like intelligence emerge outside of humans
          • 4y ago
  • api 4y ago
    We use tons and tons of pharmaceuticals whose mechanism of action is poorly understood at best or in a few cases not at all. We are even able to predict whether other compounds might work without knowing why just based on structural similarity. It’s ideal to know the full mechanism and it obviously aids engineering but there’s no reason you have to wait for that to use something. People used fire for millennia before oxidation reduction reactions were understood.
    • timid_oshima 4y ago
      In this case, it’s not so much the mechanisms of pharma we wouldn’t know, but what it is we’re even trying to cure - not having a definition of the disease or it’s symptoms and yet trying to engineer a cure would be pretty insane.
      • api 4y ago
        We have a very clear objective for a lot of this AI stuff. For the rest like the Stable Diffusion stuff it’s play.
        • timid_oshima 4y ago
          Where could I find a good definition of intelligence? At least the intelligence aimed for when we talk about general ai?
          • JieJie 4y ago
            In response to your questions, ChatGPT says, "Intelligence is the ability to process information, think abstractly, and learn from experiences. In the context of artificial intelligence (AI), intelligence typically refers to the ability of a machine or system to perform tasks that typically require human-like thinking, such as learning, problem solving, and decision making. There is no one specific definition of intelligence that is universally accepted, and different researchers and practitioners may have slightly different interpretations of what intelligence means in the context of AI." Seems a bit circular. "What's a good definition of intelligence?" "Human intelligence." "What's human intelligence?" "What we want AI intelligence to be like."
  • naasking 4y ago
    Not really. Build it, test it, notice it fails to meet what we expect of intelligence under conditions X, tweak it to fix that failure and repeat until we can't find any further failures. Then we'll have a formal model of intelligence that counts as a definition.
    • timid_oshima 4y ago
      How can we expect anything of intelligence (the criteria for testing) if we can’t define it to begin with?
      • naasking 4y ago
        Exactly the way we've been doing it: compare what it produces against known intelligence to find cases where it fails.
        • timid_oshima 4y ago
          What we’ve been doing has produced a lot of impressive stuff, but not intelligence, has it? Again, hard to tell without a good definition.
          • naasking 4y ago
            And we've learned a lot about what is and is not intelligence, and if we keep it up we'll steadily carve out the space that defines it.
          • mannykannot 4y ago
            It's not hard to tell that it has not produced human-level intelligence, so the process outlined by naasking has not yet run into an insurmountable problem.
  • nb3423 4y ago
    Just imagine if someone actually have success and an AGI just start happily chatting about everything, properly solving general problems like a 30 something engineer, mathematician or whatever. No one in the planet would know how to be sure it's not just emulating human behavior.
• dinkumthinkum 4y ago
  The basic problem is that it seems pretty clear that human intelligence is not anything like a Turing machine and no one has presented any computational system not equivalent to a Turing machine. Some might conclude this is a foundational problem.
  • mannykannot 4y ago
    It does not follow, merely from noting that there are differences between human intelligence and a Turing machine, that no "merely" Turing-equivalent device could display human-level intelligence. Every attempt I have seen so far to carry through that argument either ends up begging the question or becoming an argument from incredulity. Note that this is not an argument that it is possible, which certainly has not been conclusively established.
    • dinkumthinkum 4y ago
      I think it is on those making the claim that a system equivalent to a Turing machine can have human level or superior intelligence. Humans are vastly more intelligent than any other organism we have encountered. I may be incredulous at the idea but the scale of difference between human intelligence and programs based on Turing machines is enormous.
      • inkblotuniverse 4y ago
        The wholw point of turing machines is that they can emulate any computable system. You could have your a turing machine simulate every atom inside a human, and then you trivially have a turing machine that shows intelligence equivalent to a human.